My Thesis: Discovering Gameplay By Iterative Removal of the Visual Component

I thought it'd be nice to post some of the findings of my thesis, now that I've graduated. Originally, it was supposed to be just one game, Jojo's Adventure. But my ideas expanded, and new games came out of them. My final thesis ended up being an critical analysis of the 3 games I created, and how I identified problem areas. I will write about my process and conflicts developing the first 2...

You can see some gameplay video of Jojo and Dual Perspective here. You can also play the latest version of Dual Perspective here.

My original idea was tying a switching mechanic to the direction of a character in a 2.5D side-scroller. When the player looked left some platforms would appear while others would disappear, and vice-versa. The gameplay would hinge on the player's ability to remember the platforms that were hidden and to switch between them to navigate the level. The camera would follow the player around the level until the goal was reached. It was a relatively simple idea, and it was my first "real" attempt at a full game. Everybody thought the idea for the mechanic was sound and I was off to start production.

Initial block-world

Concept Paint Over

Fully Modeled

All this took a LONG time. About 8 months, 6 of which were almost entirely Art Production. I had underestimated how long it would take to produce all the assets. But I forged ahead. Because I thought I had a cool idea. Everybody commented on the visuals and I thought I had a great game. Then I took a new class they had added recently, called Rapid Game Development (RGD).

In RGD, the goal was to create 1 game prototype every week. In the beginning it was very difficult. But it really helped understand exactly what my production capabilities were and to come up with game ideas that were achievable in 1 week (without losing hours of sleep). It also introduced something that I hadn't considered very well: Rapid Iteration and Play Testing. It quickly became obvious that many games just weren't that good on their first shot. They were diamonds in the rough, but based off the play testing feedback, the games could be modified into something better.

Where does Jojo fit in this? I would ask people to play it, and many just wouldn't enjoy it. The visuals would attract them, but I'd see them get stuck and be frustrated in many parts.

During this class, one of the games I created was similar to Jojo. Dual Perspective (DP), borrowed Jojo's directional switching mechanic, but I made some changes. The visuals were significantly lowered, only 2 flat colors. In Jojo, if the player is standing on a platform when he turns, he will fall down. In DP, the platform will just be invisible. One more change, the camera was stationary and orthographic 2D, not 2.5D. During play testing, I noticed people's reaction was very different from Jojo. Players loved DP. It took them 5 seconds and most were smiling. Why didn't Jojo have that same reaction?

I had to analyze exactly what the memory requirements were for Jojo and DP.

Memory was a key aspect of both games, so I analyzed what the player was required to memorize in both games.

In Jojo, the camera follows the player movement. There are considerably more art assets on-screen as well as the platforms themselves. The platforms were simply cube shapes with some cartoony wonk. However, when the platforms would disappear the burden of memory on the player was: Location of a detailed 3D object in a 2D plane with a moving camera, perspective changes, as well as quick timing for successful completion. It was too complicated and was the primary reason for players getting frustrated.

In contrast to DP's memory analysis: the camera does not move, there is no perspective change (because it's flat 2D), there are no detailed environments distracting the player's attention, and platforms are simply colored rectangles. The player only has to remember the rectangle's position on the screen, because the camera never moves.


Jojo's gameplay ends up feeling twitchy, tedious and frustrating as a result of its busy environments and high memory requirement on the player.

DP's gameplay is calm. Even if there were many objects flicking on and off in visibility, they are stationary rectangles that have a low memory requirement on the player.

I learned a great deal through these projects. Exploring gameplay possibilities with memory is a tricky task. People are actually really bad at remembering things. Even more so when you add a puzzle on top. Breaking down and identifying the good parts of a bad game is incredibly important. Equally as important is to analyze why it is bad, and what can be done to fix it.

Thanks for reading!